PSYCHOLINGUISTIC

ISSUES IN
GRAMMATICAL THEORY

2017

BILLY JOSHUA LAMPUS Group 6

FADILA MARIANA KINDANGEN
Class E / VII
SWINGLY STEIV TARUMINGKENG

ENGLISH EDUCATION DEPARTMENT

FACULTY OF LANGUAGE AND ART
MANADO STATE UNIVERSITY
 ISSUES IN GRAMMATICAL THEORY

Much of what we have discussed to this point constitutes a consensus of current

thinking about linguistic concepts. In addition, linguistics has a number of issues that are
actively debated. We will discuss several of them in this section.

Psychological Reality of Grammar

As indicated earlier, much psycholinguistic research in the early and mid-1960s was
based on transformational grammar. This research was guided by the belief that the structures
and rules of transformational grammar were psychologically real; that is, that they were a part
of how people comprehend and produce language. One assumption that was made was that
the surface structure was the starting point for comprehension and that the deep structure was
the end point; the roles were assumed to be reversed for production. If so, then it would be
reasonable to assume that the distance between surface and deep structure (as measured
by the number of transformations in a sentence’s derivation) would be an accurate
index of the psychological complexity of the sentence. This view was called the deriva-
tional theory of complexity, or DTC. Early studies were encouraging. A variety of studies
showed that negative sentences such as

A. The sun is not shining.

were more difficult to comprehend than the corresponding affirmative form such as

B. The sun is shining.

But these sentences differ in meaning as well as transformational complexity, so this is hardly
conclusive. Later studies directly contradicted DTC. Sentence C is, for example,
transformationally more complex than D :

C The boy was bitten.

D.The boy was bitten by the wolf.

In transformational theory, C requires a transformation that deletes the phrase by the

wolf, so DTC would predict it would be more difficult to comprehend than D. However,
neither intuition nor experiment has revealed any relationship to processing difficulty.
Similarly, there is no psychological difference between sentences that have undergone
particle-movement transformation and those that have not. These studies have been reviewed
extensively else where (Cairns & Cairns, 1976; Fodor, Bever, & Garrett, 1974; Slobin, 1971).
As Berwick and Weinberg (1983) point out, however, these results do not necessarily mean
that the linguistic theory of transformational grammar is faulty. It could be that the linguistic
theory is correct but that some of the psychological assumptions guiding DTC are faulty.
More recent work has been more favorable to the hypothesis that linguistic theory has
psychological reality. Consider this sentence:

E. The nurse who was stationed on the seventh floor invited the chauffeur to go dancing this
evening.

This sentence contains a relative clause (who was stationed on the seventh floor),
which is a clause that begins with a wh-word and modifies a noun. Notice that the relative
clause interrupts the flow of the sentence; in particular, it introduces a gap between the noun
nurse and the verb invited. According to recent grammatical theory, relative clauses result in
the movement of the noun, thereby creating this gap. However, it is assumed that the moved
constituent leaves a trace at its earlier location, so that the presumed linguistic representation
of E would be more like F :

F. The nurse who was stationed on the seventh floor [trace] invited the chauffeur to go
dancing this evening.

If this proposal has psychological reality, then the hypothesis would be that
comprehenders would be likely to reactivate the moved noun (nurse) when its trace was
encountered. These studies have suggested that some psychologists may have overreacted to
the problems with DTC. When we see a combination of the right linguistic theory and the
right psychological experiment, better results are obtained.
Lexical versus Structural Approaches

There have long been controversies within linguistics regarding the proper way to
characterize linguistic knowledge. As we have seen, phrase-structure rules are in sufficient in
themselves to account for our linguistic capacities, and these insufficiencies led Chomsky to
propose transformational grammar.

In the years since transformational grammar was formulated, it has gone through a
number of changes. In the most recent version, Chomsky has eliminated many of the
transformational rules in previous versions of the grammar and replaced them with broader
rules, such as a rule that moves on econstituent from on elocation to another. It was just this
kind of rule on which the trace studies were based. Although newer versions of the theory
differ in several respects from the original, at a deeper level they share the idea that
grammatical rules should be structure- dependent; that is, the rules apply to constituents
rather than individual words. In a sense, transformational grammar (and its descendants) take
the grammatical constituent as the most basic unit of analysis. An alternative approach is
supplied by lexical theories of grammar. In lexical theories (for example, Bresnan, 1978),
greater emphasis is placed on individual lexical items (words) than is given in more structural
theories. This view has been influential in recent years in diverse areas of psycholinguistics,
including language comprehension, language production, and language development. Let us
go through an example to contrast structural and lexical views. In most grammars, the lexical
entry for a word includes its meaning, its spelling, its pronun- ciation, and syntactic
characteristics such as part of speech. In Bresnan’s (1978) lexical-functional grammar, lexical
entries also includes the various forms of the word (for example, kiss, kissed, kissing) and the
different kinds of sentences into which each form would fit. For verbs, this includes the
arguments or semantic roles, such as the agent (the person doing the action) and the patient
(the one to whom the action is done) that are associated with the verb as well as the surface
structure designation, such as subject or object, that goes with it.

Consider sentences G and H :

G. Mary kissed John.

H. John was kissed by Mary.

The lexical entry for kiss would indicate its underlying semantic structure as
kiss: (agent, patient)

That is, the verb requires both an agent and a patient (*John kissed is not a grammatical
sentence). In addition, the entry includes various forms of the word, including

kiss: agent = subject, patient = object

and

(be) kiss: agent = object; patient = subject

The first verb form, used in sentences in the active voice, assigns the agent role to the
surface-structure subject, and the patient to the surface object. The second form, used in
passive sentences, assigns the patient to the subject, and the agent to the object of the
preposition by. By storing this additional information in the lexical entry, the derivation of
passive sentences becomes shorter than in traditional transformational grammar. When the
surface structure includes a form of the verb kiss, that lexical entry is retrieved and fitted into
the sentence. The grammatical information in the entry allows us to interpret the sentence
semantically (that is, to interpret John aspatient). The constituent structure of a passive
sentence in lexical-functional grammar looks like a passive sentence, not like an active
sentence, and no passive transformational rule is involved. The meaning relation between
these two sentences is preserved through lexical rules that specify the relation between
different forms of a word, not by transformational rules. The major significance of lexical-
functional grammar is the shunting of most of the explanatory burden onto the lexicon and
away from transformational rules. This makes a good deal of psychological sense.
Cognitively speaking, the retrieval of items from our mental dictionary is relatively easy. In
contrast, working our way through a syntactic structure is more difficult. By storing syntactic
information in the lexical entry in the mental dictionary, lexical theories simplify the process
of comprehending sentences. This seems to provide a potentially more plausible explanation
for the nearly effortless manner in which we comprehend sentences in our everyday life.

Is Language Innate?

Another issue that has prompted considerable debate is the question of whether some of
our linguistic capacities are innate. At one level, it is obvious that experience plays a major
role in language acquisition. We all learn the language to which we are exposed, not some
other language from across the globe. Some evidence in support of the nativist view has come
from children with limited linguistic experience. In certain situations in which the child is not
pre- sented with any consistent linguistic model, they appear to have the capacity to invent
some aspects of language. This has been seen in deaf children whoseparents did not believe
in or teach ASL (Goldin-Meadow, 1982). Despite the lack of either speech or sign, these deaf
children invented a form of gestural language that was similar in some respects to ASL. They
could not have acquired this system from their parents, since the children’s facility with sign
exceeded that of their parents. Bickerton (1983) presents similar conclusions based on studies
of immigrants and their children. What kinds of linguistic capacities might be inborn?
Current thinking centers on the concept of parameters. A parameter is a grammatical feature
that can be set to any of several values. For example, the null-subject parameter deals with
whether a language permits constructions that have no subject. This parameter has two
values: null subject (the language allows sentences without a subject) or subject (the language
requires subjects for sentences to be grammatical). For example, sentence I is not
grammatical in English, but it would be in Italian or Spanish. Thus, Italian is a null-subject
language, and English is a subject language.

I. want more apples

Parameter-setting theorists (for example, Hyams, 1986), then, suggest that children are
born with the parameters and with the values of the parameters. What they must learn, from
experience, is which value is present in their native language. A rough analogy is thinking of
two restaurants. Restaurant A provides custom- ers with a small array of choices within a few
well-understood categories (that is, baked potato or fries or rice; French or Italian or ranch
dressing). Restaurant B provides customers with a large number of choices within an equally
large number of categories. Most dinnergoers would find Restaurant B informationally over-
whelming; in contrast, it would be far easier to learn what choices to make in Restaurant A.
The analogy is not perfect: we have acquired the categories in Restaurant A from experience,
whereas the language parameters are presumed to be innate. Nonetheless, there is a
fundamental similarity. Parameter-setting theorists would suggest that without built-in
categories (and values), the child would be lost in a sea of linguistic details and would not be
able to acquire a language as well as most children do. Parameter-setting models appear to
offer a tidy solution to the question ofhow innate processes interact with the child’s language
experience. Some scholars believe that the parameter-setting account is too tidy and have
pointed flaws in the model (Bloom, 1990; Valian, 1990). Nonetheless, the approach has
raised some important issues regarding the role of innate linguistic mechanisms in language
acquisition. We will discuss these issues further in Chapter 12.

Summary

Several controversial issues in grammatical theory have been discussed. One is

whether linguistic principles have psychological reality. Although researchontransformational
grammar in the 1960s suggested a negative answer, more recent re- search has reopened the
question. A second issue is whether our grammatical knowledge is better described in
structural or lexical terms. Finally, we have briefly considered whether our linguistic
knowledge may be innate.